Bigris a Sulfide Sensor That Regulates a Sulfur Transferase

Home , ETHE1, Sulfite oxidase, Sulfur dioxygenase

www.nature.com/scientificreports

OPEN BigR is a sulfde sensor that regulates a sulfur transferase/ dioxygenase required for aerobic Received: 11 December 2017 Accepted: 13 February 2018 respiration of plant bacteria under Published: xx xx xxxx sulfde stress Nayara Patricia Vieira de Lira1, Bianca Alves Pauletti1, Ana Carolina Marques2, Carlos Alberto Perez 3, Raquel Caserta4, Alessandra Alves de Souza4, Aníbal Eugênio Vercesi2, Adriana Franco Paes Leme1 & Celso Eduardo Benedetti 1

To cope with toxic levels of H2S, the plant pathogens Xylella fastidiosa and Agrobacterium tumefaciens employ the bigR operon to oxidize H2S into sulfte. The bigR operon is regulated by the transcriptional repressor BigR and it encodes a bifunctional sulfur transferase (ST) and sulfur dioxygenase (SDO)

enzyme, Blh, required for H2S oxidation and bacterial growth under hypoxia. However, how Blh operates to enhance bacterial survival under hypoxia and how BigR is deactivated to derepress operon transcription is unknown. Here, we show that the ST and SDO activities of Blh are in vitro coupled and necessary to oxidize sulfde into sulfte, and that Blh is critical to maintain the oxygen fux during A.

tumefaciens respiration when oxygen becomes limited to cells. We also show that H2S and polysulfdes inactivate BigR leading to operon transcription. Moreover, we show that sulfte, which is produced by Blh in the ST and SDO reactions, is toxic to Citrus sinensis and that X. fastidiosa-infected plants accumulate sulfte and higher transcript levels of sulfte detoxifcation enzymes, suggesting that they

are under sulfte stress. These results indicate that BigR acts as a sulfde sensor in the H2S oxidation mechanism that allows pathogens to colonize plant tissues where oxygen is a limiting factor.

Hydrogen sulfde (H2S) is a gaseous molecule that is well known for its strong odor and toxicity. In the past few years, however, H2S has gained the status of a biological efector molecule in higher organisms, since it is produced in a regulated manner by many cell types and acts as a signaling molecule in a variety of physiological processes, both in mammals and plants1–3. 4 Microorganisms, including bacteria, also produce H2S . Sulfur-reducing bacteria, for instance, can reduce 5 sulfur compounds into H2S to produce energy under anaerobic conditions . Other bacterial species, on the other 6 hand, can also generate H2S from the protein catabolism and degradation of organic matter . Despite its benefcial role in preventing oxidative stress, H2S can become toxic to both mitochondria and bacterial cells, primarily because it blocks aerobic respiration through inhibition of the cytochrome c oxidase7–9. To circumvent such problem, bacterial cells, plant and animal mitochondria have evolved common mechanisms 10–16 to eliminate H2S . In both plant and animal mitochondria, as well as in some bacterial species, three enzymes, including the Sulfde Quinone Oxidoreductase (SQOR), Tiosulfate Sulfur Transferase (TST) and Sulfur Dioxygenase (SDO) 10–12,14–16 catalyze the oxidization of H2S into sulfte . In the frst enzymatic reaction catalyzed by SQOR, H2S is complexed with sulfte to form thiosulfate. Te sulfane sulfur of thiosulfate is subsequently transferred to reduced

1Brazilian Biosciences National Laboratory (LNBio), Brazilian Center for Research in Energy and Materials (CNPEM), CEP 13083-100, Campinas, SP, Brazil. 2Department of Clinical Pathology, Faculty of Medical Sciences, State University of Campinas, 13083-887, Campinas, SP, Brazil. 3Brazilian Synchrotron Light Laboratory (LNLS), Brazilian Center for Research in Energy and Materials (CNPEM), CEP 13083-100, Campinas, SP, Brazil. 4Agronomic Institute of Campinas, Citriculture Research Center ‘Sylvio Moreira’, CEP 13490-970, Cordeirópolis, SP, Brazil. Correspondence and requests for materials should be addressed to C.E.B. (email: [email protected])

SCiENtifiC Reports | (2018)8:3508 | DOI:10.1038/s41598-018-21974-x 1 www.nature.com/scientificreports/

glutathione (GSH) by the action of TST to form glutathione persulfde (GSSH), which is then oxidized into sulfte by the action of SDO, regenerating GSH. In human mitochondria, the malfunctioning of SDO, known as ETHE1, is the cause of Ethylmalonic Encephalopathy syndrome12,17. Because sulfte is also toxic to cells, animal mitochondria further oxidizes it into sulfate, whereas in plant pathogenic bacteria such as Xylella fastidiosa and Agrobacterium tumefaciens, sulfte is secreted to the medium by a sulfte exporter13,14,18,19. We have shown that this mechanism of H2S detoxifcation in X. fastidiosa and A. tumefaciens, the causal agents of the Citrus Variegated Chlorosis (CVC) and Crown Gall disease, respectively, is encoded by the bigR operon. Te bigR operon is regulated by BigR (bioflm growth-associated repressor), a winged helix-turn-helix repressor that recognizes the −10 region of the operon, blocking its transcription13,20. Te repressor role of BigR is governed by the redox status of two conserved cysteines (Cys42 and Cys108) which, when oxidized into a disulfde bond, alter the conformation of the winged-helix DNA-binding region of the repressor, leading to repressor-DNA dissociation and operon activation13. In addition to BigR, the bigR operon encodes a two-domain DUF442-ETHE1 protein, originally named Blh (β-lactamase-like hydrolase), and a TauE-related sulfte transporter13,20. Protein sequence alignments and structural modeling studies suggested that the Blh DUF442 domain might function as a rhodanese or sulfur transferase (ST), whereas its β-lactamase domain would function as an SDO similar to human ETHE113. In fact, genetic studies showed that Blh becomes critical to bacterial growth under hypoxia or under conditions where H2S is produced13. Tis is particularly relevant for aerobic organisms like X. fastidiosa and A. tumefaciens which colonize plant tissues where the O2 tension is low. In such environment, H2S accumulation could potentially inhibit bacterial respiration. Although the lack of Blh impairs bacterial growth under oxygen-limited conditions13, how exactly Blh enhances bacterial survival under hypoxia or H2S stress was unclear. In addition, how BigR is deactivated to derepress operon transcription was also unknown13. Here we show that the DUF442 and ETHE1 domains of Blh display ST and SDO activities upon GSH and GSSH, respectively, producing sulfte, and that such enzymatic activities are in vitro coupled and required to maintain the O2 fux during bacterial respiration when O2 becomes limited to cells. Our data also reveal that H2S and polysulfdes react with Cys108 and induce the Cys42-Cys108 disulfde bond formation in BigR, leading to repressor inactivation and operon transcription. We also present evidence suggesting that citrus plants infected with X. fastidiosa are under sulfte stress. Results The ETHE1 domain of Blh is an iron-containing sulfur dioxygenase. In our previous work, we provided evidence that the ETHE1 domain of Blh would function as an SDO like the mitochondrial ETHE1 13 protein, since the blh-defective mutant of A. tumefaciens accumulated higher levels of H2S . In addition, based on structural modeling studies, we hypothesized that the Blh ETHE1 domain would use GSSH as a substrate13. To verify these assumptions, the recombinant ETHE1 domain of Xylella Blh, corresponding to residues 143 to 407, was purifed and tested for the SDO activity upon GSSH. In agreement with the predictions, we found that O2 is consumed in the presence of GSSH and the Blh ETHE1 domain, and that sulfte is generated as a reaction product (Fig. 1a and b). Typically, ~1.3 ± 0,3 µmol O2 was consumed per mg of the ETHE1 domain per min at 25 °C (Fig. 1c). To better characterize the kinetic properties of the Blh ETHE1 domain, the SDO reaction was performed in the presence of increased amounts of GSSH. Te results indicate that the ETHE1 domain shows an ordinary Michaelis-Menten kinetics with estimated apparent Km and Vmax of approximately 0.18 mM GSSH and 3.4 µmol −1 −1 O2 mg min , respectively (Fig. 1d). Tese kinetics parameters are comparable to those of other characterized SDOs11,15. Our molecular modeling studies also suggested that a metal ion, possibly iron, could be part of the active site of the Blh ETHE1 domain13. Consistent with this, X-ray fuorescence analysis revealed higher levels of iron in the ETHE1 domain sample, relative to bufer alone or BSA (bovine serum albumin), used as negative control samples (Fig. 1e). Tis result is in line with the fact that iron was also found to bind the Blh-related CstB and human ETHE1 enzymes15,21.

The Blh DUF442 domain is a sulfur transferase. Protein sequence alignment and molecular modeling studies suggested that despite sharing the same fold of protein tyrosine phosphatases, the Blh DUF442 domain was not only structurally similar to the catalytic domain of the E. coli rhodanese YnjE22, but also presented a conserved rhodanese active site13. To further investigate this, the recombinant DUF442 domain of Xylella Blh (residues 1 to 141) was purifed and tested for a rhodanese or sulfur transferase (ST) activity using thiosulfate as a sulfur donor and cyanide as a sulfur acceptor23. Te results show that the DUF442 domain of Blh was sufcient to catalyze the transfer of the sulfane sulfur of thiosulfate to cyanide, generating thiocyanate (SCN) and sulfte as the reaction products (Fig. 2a and b). Te recombinant full-length Blh also showed ST activity upon thiosulfate and cyanide which was, on average, twice of that exhibited by the DUF442 domain alone (Fig. 2a). Because the DUF442 domain was relatively less soluble and active than the full-length Blh when expressed in E. coli cells, the kinetics properties of the DUF442 domain was determined using the full-length Blh protein. Te −1 −1 apparent Km and Vmax values for the Blh ST activity were 6.2 mM thiosulfate and 18.1 µmol SCN mg min , 15 respectively (Fig. 2c), which are also comparable to the Km and Vmax values determined for the S. aureus CstB . Tese data thus show that the DUF442 domain of Blh functions as a sulfur transferase and, therefore, will be referred here as the Blh ST domain.

The ST and SDO activities of Blh are coupled. We found that the ST domain of Blh can also use GSH as a sulfur acceptor to produce sulfte and GSSH when thiosulfate is the sulfur donor (Fig. 3a). Considering that the Blh ETHE1 domain uses GSSH as substrate and its SDO activity generates GSH, we tested whether the ST

SCiENtifiC Reports | (2018)8:3508 | DOI:10.1038/s41598-018-21974-x 2 www.nature.com/scientificreports/

Figure 1. Te ETHE1 domain of Blh is an iron-containing sulfur dioxygenase. (a) SDO activity of the Blh ETHE1 domain (black line), relative to bufer with no protein (red line), according to the reaction scheme shown. Oxygen consumption, as a measure of the SDO activity of the Blh ETHE1 domain, occurs only afer the addition of substrate GSSH to the reaction mixture (arrow). (b) Sulfte production at the terminus of the SDO reaction estimated by the sulfte test strip, according to the scale. (c) Average O2 concentration per mg of protein (ETHE1 domain) per min at 25 °C, relative to control (no protein in the reaction mixture). Values are the mean of three independent measurements and error bars indicate standard deviations. (d) SDO activity as function of the GSSH concentration, showing that the ETHE1 domain displays a Michaelis-Menten kinetics with an −1 −1 estimated apparent Km and Vmax values around 0.18 mM GSSH and 3.4 µmol O2 mg min , respectively. (e) X-ray fuorescence analysis showing higher levels of iron in the ETHE1 domain sample, relative to bufer alone (no protein) or BSA, as negative control samples.

and SDO activities of Blh were coupled. To verify this, the ST and ETHE1 domains of Blh were purifed separately and O2 consumption, as a result of the SDO activity (Fig. 1a), was measured in the reaction. Te results show that O2 consumption is not signifcantly altered when the ST or the ETHE1 domain is added alone into the reaction mixture containing thiosulfate and GSH. However, O2 consumption increases signifcantly as soon as the second Blh domain is added to the reaction mixture (Fig. 3b). When the full-length Blh is used, an instant and sharper O2 consumption is observed (Fig. 3b). Together, these results confrm that the ST and SDO activities of Blh are in vitro coupled and that GSH, consumed in the ST reaction, is restored in the SDO reaction.

Blh is essential to maintain aerobic respiration under oxygen-limited conditions. We showed previously that the bigR operon is important for H2S detoxifcation and that Blh is required to sustain bacterial 13 growth under O2-limited concentrations . Given that H2S competes with O2 for cytochrome c oxidase binding, we decided to investigate the role of Blh as a sulfde oxidation enzyme, particularly when O2 becomes limited to the cells. To test this, we performed respiration assays using the wild type and corresponding blh− and

SCiENtifiC Reports | (2018)8:3508 | DOI:10.1038/s41598-018-21974-x 3 www.nature.com/scientificreports/

Figure 2. Te Blh DUF442 domain is a sulfur transferase. (a) ST activity of the full-length Blh or its DUF442 domain alone (ST domain), relative to bufer with no protein. Te ST activity, as a measure of iron-SCN formation, used thiosulfate as the sulfur donor and cyanide as the sulfur acceptor, according to the reaction scheme shown. Values are the mean of nine independent measurements and error bars indicate the standard deviation. (b) Sulfte production at the terminus of the ST reaction estimated by the sulfte test strip, according to the scale, indicating that both the full-length Blh and DUF442 domain catalyzed the transfer of the sulfane sulfur of thiosulfate to cyanide, generating SCN and sulfte as the reaction products. (c) ST activity of full-length Blh as function of the thiosulfate concentration, showing that Blh displays a Michaelis-Menten kinetics with an −1 −1 estimated apparent Km and Vmax values of ~6.2 mM thiosulfate and 18.1 µmol SCN mg min , respectively.

bigR− deletion mutants of A. tumefaciens described previously20. Te bacterial cells were incubated in medium containing glucose and O2 consumption was monitored until its full depletion. −18 −1 −1 Te O2 fux, as a measure of the aerobic respiration, was in the order of 10 mol cell s for the three bac- 24 teria, which is consistent with the average of O2 utilization described for other cell types . We found that the − blh mutant showed a relatively lower O2 fux compared to the wild type cells under O2-replete conditions (above −1 − 20 nmol mL ), whereas the bigR mutant, on the other hand, a slightly higher O2 fux relative to the wild type − − cells (Fig. 4a, b and c). Interestingly however, the diferences in the O2 fux between the blh and bigR mutants, relative to the wild type cells, increased signifcantly as the O2 concentration in the medium approached zero − (Fig. 4d). Te O2 fux of blh cells, for instance, was approximately one tenth of that of the wild type cells at 0.2 −1 −1 nmol mL O2, but nearly the same of that of the wild type cells at 5.0 nmol mL O2 (Fig. 4d). On the other hand, − − −1 the bigR cells maintained a higher O2 fux compared to the wild type and blh cells even at 0.2 nmol O2 mL − (Fig. 4d). Tese results thus confrm that O2 consumption is afected in the blh mutant and that the bigR operon

SCiENtifiC Reports | (2018)8:3508 | DOI:10.1038/s41598-018-21974-x 4 www.nature.com/scientificreports/

Figure 3. Te couple ST and SDO activities of Blh. (a) ST activity of the Blh ST domain as a measure of sulfte production when cyanide (CN) or GSH are used as sulfur acceptors and thiosulfate as the sulfur donor. Values are the mean of three independent measurements and error bars indicate the standard deviation. In addition to sulfte, the ST activity upon GSH is expected to generate GSSH, according to the reaction scheme shown. (b) O2 consumption as a measure of the SDO activity of the Blh ETHE1 domain showing that the O2 concentration in the medium is not signifcantly altered when the ST domain is added to the reaction mixture containing thiosulfate and GSH. Nevertheless, O2 consumption increases signifcantly soon afer the addition of the ETHE1 domain to the reaction mixture (red line). Similar results are observed when the ETHE1 domain is added frst to the reaction mixture (blue line). When the full-length Blh is used, a sharper decrease in the O2 levels is observed (black line), indicating that the ST and SDO activities of Blh are in vitro coupled.

is critical to improve aerobic respiration under hypoxia, a condition experienced by Xylella and Agrobacterium cells inside plant tissues.

Sulfte is toxic to citrus cells and accumulates in CVC plants. Te results presented in Figs 1b, 2b and 3a show that sulfte is generated by both ST and SDO activities of Blh. In contrast to mitochondria, where sulfte produced by the TST and ETHE1 pathway is further oxidized to non-toxic sulfate before it is secreted14, in A. tumefaciens, sulfte is excreted to the medium13. Because sulfte is toxic to plant cells25–28, we suspected that sulfte secreted by the action of the bigR operon could become toxic to the host plant, as bacterial coloniza- tion progresses. To explore this idea, we incubated sweet orange leaf discs in the presence of increased amounts of sulfte or sulfate. We found that sulfte, but not sulfate, caused a dose-dependent chlorosis in citrus tissues, indicating that it is toxic to citrus cells (Fig. 5a). Because sulfte caused the yellowing of the citrus leaf discs, we hypothesized that it might contribute to the CVC symptoms, since typical CVC lesions are surrounded by chlorotic halos (Fig. 5b). To investigate this, we measured the sulfte levels in symptomatic and asymptomatic CVC plants, relative to non-infected (healthy) plants. Sulfte levels in non-infected citrus leaves were in the range of 150 nmol per gram of fresh weigh (Fig. 5c), which is close to what has been reported for tomato and Arabidopsis27,28. We found that the sulfte levels were signifcantly higher in both symptomatic and asymptomatic leaves, relative to non-infected plants (Fig. 5c). Consistent with these results, bacterial quantifcation in the petioles of the corresponding leaves revealed high bacterial titles in both symptomatic and asymptomatic leaves (Fig. 5d). To further investigate whether Xylella-infected leaves were under sulfte stress, the mRNA levels for Sulfte Reductase (SiR) and Sulfite Oxidase (SO), the principal enzymes responsible for sulfite detoxification in plants26–28, were examined. We found that both SiR and SO are signifcantly increased in leaves infected with Xylella, compared to non-infected leaves (Fig. 5e). Tus, although a clear link between sulfte levels and visible CVC symptoms could not be established, the results show a correlation between increased sulfte levels and bacterial cell population with the expression of sulfte detoxifcation genes in Xylella-infected plants, indicating that plants with CVC are under sulfte stress.

H2S and polysulfdes oxidize BigR and activate the bigR operon. BigR has been suggested to function as a redox sensor in the sense that it loses afnity to DNA, allowing transcription of the operon, when its Cys42 and Cys108 residues form an intra-chain disulfde bond13. To investigate what molecules might induce BigR oxidation, we performed gel-shif assays in the presence of GSSH, since GSSH and other persulfdes were recently shown to regulate the bigR-related cst operon in S. aureus and the sulfde-responsive regulator SqrR in Rhodobacter capsulatus15,29–31. Notably, we found that although GSSH abolished the BigR binding to its target DNA, the control reaction containing only the elemental sulfur also disrupted the BigR-DNA complex (Fig. 6a). Tis result strongly suggested that sulfde or polysulfdes, reported to be present in sulfur-containing compounds, was responsible for BigR oxidization. Indeed, we found that the elemental sulfur source used in our experiments contained polysulfdes, as revealed by the characteristic yellow pigmentation in alkaline solution32 (Fig. 6b). Te presence of polysulfdes in the elemental sulfur source was also evidenced by the absorbance curve with

SCiENtifiC Reports | (2018)8:3508 | DOI:10.1038/s41598-018-21974-x 5 www.nature.com/scientificreports/

Figure 4. Blh is required to maintain aerobic respiration under O2-limited conditions. O2 consumption (black lines) and O2 fux (red lines), as a measure of aerobic respiration, exhibited by the wild type (a) A. tumefaciens and corresponding blh− (b) and bigR− (c) cells in medium containing glucose. Te blh− cells showed a lower −1 O2 fux (b) compared to the wild type cells (a) under O2-replete conditions (above 20 nmol mL ), whereas − the bigR cells (c) showed a higher O2 fux relative to the wild type cells. (d) O2 fux as a function of the O2 − − concentration showing that the diferences in the O2 fux between the blh and bigR mutants, relative to the wild type cells, increased signifcantly as the O2 concentration in the medium approaches zero. Te O2 fux of − −1 blh cells at 0.2 nmol mL O2 was approximately 10 × lower than that of the wild type cells, but nearly the same −1 − of that of the wild type cells at 5.0 nmol mL O2. Conversely, the bigR cells maintained a high O2 fux even at −1 − 0.2 nmol mL O2, compared to the wild type and blh cells.

maximum peaks at 300 and 375 nm33 (Fig. 6c). Moreover, our elemental sulfur source also reacted with bismuth ion at alkaline pH forming a brownish-orange pigment, a strong indication of the presence of H2S or the hydrosulfde ion (HS−) in solution (Fig. 6b). Tese results, together with the fact that polysulfdes have recently 33 been shown to link H2S to protein thiol oxidization , led us to investigate whether BigR could directly sense the H2S gas. To test this, recombinant BigR was incubated with gaseous H2S in a stainless-steel jar for 30 min, and the protein was subjected to gel-shif assays. We found that BigR treated with H2S no longer bound the target DNA (Fig. 6d), indicating that H2S itself can lead to BigR Cys42-Cys108 disulfde bond formation and operon activation. Because gaseous H2S could not be employed to treat bacterial cells in ordinary bacterial growth chambers, we instead tested whether our sulfur preparations containing HS− and/or polysulfdes could activate the bigR operon. Tus, A. tumefaciens cells carrying a GFP reporter gene driven by the bigR operon promoter20 were grown in the presence or absence of HS−/polysulfdes. We found that the reporter cells exposed to the HS−/polysulfde solution showed increased GFP fuorescence in a dose-dependent manner (Fig. 6e), indicating that HS− and/or polysulfdes were responsible for the operon activation, and like the H2S gas, were inducing the disulfde bond − formation in BigR. To confrm this, recombinant BigR that had been incubated with gaseous H2S or the HS / polysulfde solution was subject to mass spectrometry analysis. Te results show that treatments of BigR with − the H2S gas or HS /polysulfde solution led to the incorporation of a ‘SH’ or ‘SSH’ group to the sulfur atom of Cys108, forming cysteine persulfdes (Supplementary Table S1). Moreover, the peptide that shows the disulfde 13 bond involving Cys42 and Cys108 was also detected in BigR samples treated with gaseous H2S (Supplementary Fig. S1). Tus, considering that cysteine persulfdes can serve as intermediates of disulfde bond formation when 30,33 a second cysteine is nearby , the results show that BigR senses sulfde species and that H2S itself is the actual modulator of BigR in vivo.

BigR is modifed by sulfenic acid and glutathionylation at regulatory Cys residues. Sulfenic acid at cysteine residues has been reported to serve as an intermediate in disulfde bond formation34–36. In addition, 37 H2S has been shown to react with sulfenic acid to form persulfdes . In the BigR-related transcriptional regulator HlyU from Vibrio cholera, Cys38 is modifed to cysteine sulfenic acid38. Because this cysteine is structurally equivalent to Cys42 in BigR, we tested whether Cys42 in BigR could also be modifed to cysteinyl sulfenic acid.

SCiENtifiC Reports | (2018)8:3508 | DOI:10.1038/s41598-018-21974-x 6 www.nature.com/scientificreports/

Figure 5. Xylella-infected plants are under sulfte stress. (a) Leaf discs of Pineapple sweet orange incubated with increased amounts of sulfte or sulfate, showing that sulfte, but not sulfate, induces chlorosis in citrus. (b) Example of a symptomatic CVC leaf of Pineapple plants showing typical CVC lesions surrounded by chlorotic halos (lef panel), compared with an asymptomatic leaf (right panel). (c) Sulfte levels measured in symptomatic (S) and asymptomatic (A) CVC leaves showing higher sulfte levels in symptomatic and asymptomatic leaves relative to non-infected control (C) plants. Values are the means of six independent measurements and error bars indicate the standard deviation, whereas asterisks indicate statistically signifcant diference between the means. (d) Quantitative PCR analysis showing high bacterial titles in both symptomatic and asymptomatic leaves of CVC plants. No PCR amplifcation is detected in non-infected plants. (e) Quantitative PCR analysis showing the relative mRNA levels of SiR and SO between symptomatic and asymptomatic CVC leaves, compared to non-infected control plants. Values in ‘d’ and ‘e’ are the means of fve independent biological replicates and asterisks indicate statistically signifcant diference between the means.

Using dimedone, which specifcally reacts with sulfenic acid36, we found that not only Cys42 but also Cys108 is modifed to cysteine sulfenic acid (Table S1). Moreover, we found that Cys42 and Cys108 can also be linked to GSH (Table S1). Since GSH can react with cysteinyl sulfenic acid34, our results indicate that Cys42 and Cys108 in BigR can be modifed to cysteinyl sulfenic acid to subsequently react with GSH or SH−. Discussion In our previous work, we showed that the bigR operon was required for H2S detoxifcation to allow bacterial 13 growth under O2-limited conditions and suggested that Blh was implicated in the oxidization of H2S into sulfte . Moreover, we predicted that the DUF442 and ETHE1 domains of Blh would display independent but coupled enzymatic activities and that DUF442 was structurally related to bacterial rhodaneses13. Here, we confrmed that the SDO and ST/rhodanese activities of the ETHE1 and DUF442 domains of Blh are in vitro coupled, corroborating data that show that such H2S oxidation mechanism is conserved in bacteria, plant and animal mitochondria10–12,14–16. Our data also unveiled the fundamental question of why Blh is critical for bacterial growth under hypoxia. Since H2S blocks aerobic respiration by competing with O2 for cytochrome C oxidase binding, Blh helps to maintain a high O2 fux, especially when O2 becomes a limiting factor to the cells. Importantly, a comparative genomics study indicated that strains of X. fastidiosa possess a simple respiratory complex with just one terminal oxidase that would allow aerobic respiration only at high O2 levels, as they seem to lack cytochromes with high O2 afn- ity39. Tus, for aerobic organisms like X. fastidiosa and A. tumefaciens, the bigR operon becomes essential for survival, since it allows these pathogens to respire inside xylem vessels and roots where the O2 concentration can be very low40–42. Because sulfte is produced by the ST and SDO reactions catalyzed by Blh, and is toxic to plant cells26–28, we decided to investigate whether sulfte could contribute to CVC symptoms, since it also induced chlorosis in citrus leaves (Fig. 5a). Although there was no apparent direct correlation between increased sulfte levels and typical chlorosis in CVC symptomatic leaves (Fig. 5), we found a correlation between increased sulfte levels and Xylella population with the augmented expression of Sir and SO. Tus, although further studies will still be necessary

SCiENtifiC Reports | (2018)8:3508 | DOI:10.1038/s41598-018-21974-x 7 www.nature.com/scientificreports/

Figure 6. H2S and polysulfde oxidizes BigR and activates the bigR operon. (a) Gel-shif assay stained with ethidium bromide showing formation of the BigR-DNA complex (arrow) in the control mix reaction (C), or in the presence of GSH. Te BigR-DNA complex is not only abolished in the presence of GSSH, but also in the presence of acetonic sulfur (S), used in combination with GSH to generate GSSH. FP designates the free probe. (b) Detection of HS− and polysulfdes in the commercial source of elemental sulfur. Acetone adjusted to pH 12 with potassium hydroxide was used as a control solution (C). Addition of sulfur (S) caused the solution to turn yellow indicating the presence of polysulfdes in the sample. Addition of bismuth citrate (Bi) to the acetone solution formed a white precipitate which turned into an orange-brownish color with sulfur addition (Bi + S), indicating the presence of HS− in the sample. (c) Absorbance curve of the ammonium sulfde solution (60 mM) in the presence and absence of sulfur showing peaks at 300 and 375 nm characteristic of polysulfdes. (d) Gel- shif assay showing that the BigR protein treated with H2S gas no longer binds DNA. (e) Fluorescence images of A. tumefaciens cells carrying the GFP reporter plasmid for the bigR operon growing in the presence and absence of the polysulfde solution at 0.2%, 2.0% or 10% v/v. Cells exposed to the polysulfde solution show increased GFP fuorescence in a dose-dependent manner, indicating that HS− and/or polysulfdes are responsible for the operon activation.

to demonstrate whether sulfte secreted by Xyllela cells would play an important role in the development of CVC symptoms, our data indicate that plants infected with Xylella are under sulfte stress. Another fundamental question that remained unanswered was regarding the molecule responsible for BigR oxidation and operon activation in vivo. Due to its redox potential, H2S was initially not considered as a candi- date13. However, in light of the work by Greiner and colleagues33 who showed that, in spite of its redox potential, H2S can lead to protein disulfde bond formation via polysulfde intermediates, we decided to investigate whether − polysulfdes or H2S could oxidize BigR and activate the operon. Here, we show that solutions containing HS and/ or polysulfdes activated the bigR operon and induced persulfde formation at Cys108. Like the H2S-induced 43 sulfydration of human Protein Tyrosine Phosphatase PTP1B , gaseous H2S fushed into solutions of BigR also induced sulfydration of Cys108 and Cys42-Cys108 disulfde bond formation in the repressor. Notably, the sulfur atom of Cys108 is more surface exposed in the BigR 3D structure, compared with Cys42, and thus more prone to react with HS− or polysulfdes to form the cysteine persulfde (Supplementary Fig. S2a). Moreover, a comparison of the crystal structures of reduced (SH) and oxidized (S-S) BigR shows that the side chain of Trp104, which sits in

SCiENtifiC Reports | (2018)8:3508 | DOI:10.1038/s41598-018-21974-x 8 www.nature.com/scientificreports/

between Cys42 and Cys108 in reduced BigR and prevents the disulfde bond formation, fips almost 90 degrees in the oxidized BigR13 (Fig. S2a and b). It is thus possible that the initial formation of a Cys108 persulfde in reduced BigR could induce a conformational change in the Trp104 side chain opening space for the two cysteines to interact (Supplementary Fig. S2b). It is also noteworthy that Cys42 and Cys108 can undergo S-glutathionylation and S-sulfenylation. Even though these cysteines modifcations were the result of in vitro reactions with recombinant BigR, it is known that cysteine sulfenic acid can react not only with GSH, but with H2S to form cysteine persulfdes, which eventually 34,37,44,45 − lead to disulfde bond formation between vicinal cysteines . Our data thus suggest that H2S/HS is likely to be the actual signaling molecule that activates the bigR operon via cysteine sulfydration and disulfde or polysulfde bond formation in BigR. Terefore, instead of acting as a simple redox sensor, as previously suggested13, BigR appears to play a role as a H2S or sulfde sensor. In this regard, despite sharing similar features to the S. aureus CstR persulfde sensor29,30, BigR is phylogenetically related to the SqrR repressor recently shown to act as a sensor 31 for H2S-derived reactive sulfur species . We hope that the data presented here indicating that BigR is a H2S sensor and that Blh is required for aerobic respiration under hypoxia will ofer new opportunities for pathogen control. Methods Cloning, expression and purifcation of Blh and corresponding DUF442 and ETHE1 domains. Te 9a5c X. fastidiosa Blh full-length gene (WP_010893291) and corresponding DUF442 and ETHE1 regions were amplifed with the oligonucleotide pairs CATATGAGAATAGTAGACATC/CTCGAGTCACCATGCGGCACC, CATATGAGAATAGTAGACATC/GAGCTCGTTAGGCTTGACGTTCTAACCAGG and CATATGCAGACTCCG CGCGTTTCAGG/CTCGAGTCACCATGCGGCACC, respectively, and subcloned into the NdeI/XhoI-SacI sites of pET28a vector for the expression of recombinant 6xHis-fusion proteins. Te plasmids were verifed by DNA sequencing and inserted into Escherichia coli BL21 pLySE cells, which were grown at 20 °C for 30 h in ZYM5052 medium46, supplemented with kanamycin (100 µg/mL), for protein production. Te cells were suspended in lysis bufer (20 mM Tris-HCl, pH 7.4, 200 mM NaCl, 300 mM ammonium sulfate, 10% glycerol) and disrupted in a French press (25 kpsi) at 4 °C. Te suspension was centrifuged and the recombinant proteins were purifed on a NiNTA column (GE-Healthcare) equilibrated with lysis bufer. Te column was washed with 10 column volumes of lysis bufer containing 50 mM imidazole and the proteins were eluted with lysis bufer containing 200 mM imidazole. Protein purity was verifed by denaturing polyacrylamide gels and quantifed on a NanoDrop. Protein yields were around 1.0, 0.6 and 4.0 mg/L for the full length Blh, ST and SDO domains, respectively.

Enzyme assays. Te SDO activity of the Blh ETHE1 domain was determined by measuring the O2 consumption in the reaction mixture, as described by Hildebrandt and Grieshaber11. Te purifed ETHE1 domain (12 µg) was incubated at 25 °C in 2 mL (fnal volume) of 100 mM potassium phosphate bufer, pH 7.4, containing 1 mM GSH. Te reaction started with the addition of 30 µL of a saturated acetonic sulfur solution, and O2 consumption was measured on a Hansatech oxygraph. Te sulfte concentration in the reaction mixture was estimated using the MERCKOQUANT sulfte test strips (Merck-Millipore, USA) at the end of the SDO reaction. For estimation of the specifc SDO activity, Vmax and Km values, the enzyme assays were performed with increasing amounts of GSSH (0 to 1.2 mM). Te amount of GSSH obtained in the GSH/saturated acetonic sulfur solution was estimated by the methylene blue method11. Te rhodanese activity of full-length Blh or its DUF442 domain alone was determined spectrophotometrically as a measure of the iron-thiocyanate formation, using thiosulfate and cyanide as sulfur donor and acceptor, respectively23. Te purifed proteins (12 µg) were incubated at 25 °C for 5 min in 12 mM potassium phosphate bufer, pH 7.4, containing 12 mM sodium thiosulfate and 12 mM potassium cyanate. Te reaction was stopped with 1.8% formalin, followed by the addition of 100 mM iron nitrate, dissolved in 14% nitric acid, and the samples were read for absorbance at 460 nm. Te amount of sulfte produced in the reaction was estimated using the sulfte test strips. For estimations of the specifc enzyme activity, Vmax and Km of full-length Blh, the rhodanese reaction was performed varying the thiosulfate concentration from 0 to 100 mM. Te ST activity of full-length Blh or of the DUF442 domain alone was determined spectrophotometrically as a measure of sulfte production using the fuchsine method47. Purifed proteins (12 µg) were incubated at 25 °C for 5 min in 50 mM acetate-bufer, pH 7.5, containing 10 mM sodium thiosulfate and 10 mM GSH. Alcoholic potassium hydroxide was added to samples at 0.16% fnal concentration. Te samples were dissolved to 1:1 (V/V) in fuchsine reagent and read at 580 nm. A standard curve for sodium sulfte concentrations ranging from 0.4 µM to 2 mM was generated to estimate the amounts of sulfte in the samples. Te coupled ST and SDO activities of Blh was determined in the oxygraph in 50 mM Tris-acetate, pH 7.4, containing 10 mM thiosulfate and 10 mM GSH, using the DUF442 and ETHE1 domains purifed separately. Te ST reaction started with the addition of 6 µg DUF442, and afer 3 min, the ETHE1 domain (6 µg) was added to the reaction mixture. Te reaction was also performed starting with the addition of the ETHE1 domain frst, or using the full-length Blh as a control. O2 consumption was measured throughout the assay.

X-ray fuorescence analysis. X-ray fuorescence measurements of the ETHE1 and BSA samples were performed at the XRF beamline of the Brazilian Synchrotron Light Laboratory (LNLS), as described previously48. A white beam of 0.1 mm high by 5 mm wide was used to excite the ETHE1 protein, or BSA, used as control, under total external refection conditions of the X-ray beam for sample excitation. Samples were prepared by pipetting 5 µL of the purifed proteins (~4 µg) onto a Perspex support. Samples dried at room temperature were measured for 300 s and the collected X-ray fuorescence spectra were evaluated using the PyMCA program49.

SCiENtifiC Reports | (2018)8:3508 | DOI:10.1038/s41598-018-21974-x 9 www.nature.com/scientificreports/

Respirometry assays. High-resolution respirometry assays were performed using an Oroboros 2 K oxygraph. Te A. tumefaciens cells were grown in YEP medium (1.0% peptone, 1.0% yeast extract, 0.5% NaCl, 1.5% agar) at 30 °C for 48 h. Bacterial colonies were suspended in AB medium50, supplemented with 2 mM sodium phosphate and 40 mM MES, pH 5.8, and the optical density at 600 nm immediately adjusted to 1.5. Two mL of the cell suspensions containing approximately 1.0 × 107 cells were analyzed in the Oroboros equipment previously calibrated with the same medium. O2 consumption was recorded at 25 °C until its full depletion.

Sulfte toxicity assay. Leaves from Citrus sinensis (Pineapple) plants, maintained in the greenhouse, were detached and washed with sterile water. Leaf discs were excised, embedded in water solutions of sodium sulfte or sodium sulfate at 0.5, 1.0, 5.0 and 10 mM concentrations, and placed onto flter paper wet with the same solutions inside petri dishes. Te leaf discs were incubated at 25 °C for 24 to 72 h under a 12 h light/12 h dark regime, afer which they were photographed.

Plant inoculations and quantitative RT-PCR. Sweet orange Pineapple plants were infected with X. fastidiosa (9a5c strain) according to Almeida et al.51. Six leaves showing typical CVC symptoms, or no visible symptoms, were collected and their petioles were used for bacterial quantifcation. Petioles were weighed, ground in liquid nitrogen and total DNA was extracted using DNeasy Blood and Tissue kit (Qiagen). Bacterial population was determined by qPCR using a standard curve, as described previously52. Amplifcations were performed using TaqMan PCR Master Mix (Applied Biosystems) with 10 pmol of primers CVC-1 (TGA AGA TCA TGC AAA AAA CAA) and CCSM-1 (GCGCATGCCAAGTCCATATTT), probe TAQ-CVC-(6FAM) AACCGCAGCAGAAGCCGCTCATC (TAMRA) (Applied Biosystems) and 270 ng of DNA. Reactions were performed in triplicates using an ABI PRISM 7500 Sequence Detector System (Applied Biosystems). Te Ct (threshold cycle) and DNA concentration values were correlated by linear regression. For SiR and SO quantification, total RNA was extracted from the same collected leaves using TRIzol (Termo Fisher Scientifc, USA). Te quantity and quality of the RNA samples were evaluated by denatured gel electrophoresis and absorbance at 260/280 nm. cDNA was synthesized using the SuperScript III Reverse Transcriptase (Termo Fisher Scientifc), following the manufacturer’s procedure. Te oligonucleotides used for the amplification of the SiR (GCCTTCGCATTCGCTCTCA/CGTACGGAACGGGAAAAGC) and SO (TCGAGGATGCCGGGACTA/ACGAGGCGGCTCTTGAGA) genes (GenBank accessions XM_006470566 and XM_006469992.2, respectively) were designed using the Applied Biosystems Primer Express 2.0 sofware, and the amplifcations were carried out using a 7500 PCR system (Applied Biosystems, USA). Te citrus actin and malate translocator genes were used as internal control for normalization53,54. Total RNAs from fve independent leaves were used as biological replicates, and three technical replicates for each biological sample were performed. Te relative gene expression levels among samples were determined using the 7000 System SDS sofware v2.3 (Applied Biosystems) with default parameters

Sulfte measurements in citrus leaves. Leaf powder (250 mg fresh weight) from the same symptomatic, asymptomatic or healthy pineapple plants used for Sir and SO mRNA quantifcations were incubated in 1 mL Tris-HCl pH 8.0 with 500 mM sodium sulfate solution for 10 min at 25 °C to avoid spontaneous conversion of sulfte into sulfate55. Te suspensions were centrifuged at 18.000 × g for 5 min and the supernatants were collected and mixed with 3% KOH diluted in ethanol (1:1 v/v), as described by Leinweber and Monty55. Samples were centrifuged at 18.000 × g for 15 min at 4 °C and the supernatants were mixed (5:1 v/v) with fuchsine solution prepared according to Leinweber and Monty47. Afer 10 min incubation at room temperature, samples were measured for absorbance at 575 nm. Te amount of sulfte in the samples was estimated using a sulfte calibration curve determined with freshly prepared sodium sulfte solutions47.

Detection of H2S and polysulfdes in elemental sulfur sources. Saturated acetonic sulfur solutions were adjusted to pH 12 with potassium hydroxide at room temperature and inspected for the yellow color pro- duction32,33. Te polysulfde solution was prepared according to Kleinjan et al.56. Elemental sulfur (0.15 g) was dissolved into 20 mM potassium phosphate bufer, pH 8.0, containing 60 mM ammonium sulfde and the solution was incubated for 16 h at 50 °C. Polysulfde formation was evaluated by absorbance at 300–490 nm33. To evaluate the presence of H2S in the elemental sulfur sources, bismuth citrate was added to the saturated acetonic sulfur solutions, which were inspected for a brownish-orange color.

Incubation of BigR with gaseous H2S. Recombinant BigR was purifed by afnity chromatography as previously described20. Aliquots of 0.3 mL of purifed BigR (0.2 mg/mL) in 50 mM Tris-HCl pH 7.4, NaCl 200 mM NaCl, 10% glycerol 10% and 200 mM imidazole, were placed inside a sealed stainless-steel jar specially designed to fow gases into the sample. Nitrogen gas at 20 psi was fushed through the system for 5 min at 1 mL/min to remove O2. Subsequently, 5.0% H2S at 20 psi was fushed at 1 mL/min through the sample for up to 1 h. Afer the treatment, protein samples were subjected to gel-shif and mass spectrometry analysis as described below.

Gel-shift assays. Gel-shif assays were performed as described previously20, except that the DNA-protein complexes were detected by ethidium bromide. Approximately 2 µg of BigR, purifed by afnity chromatography, was incubated in binding bufer (10 mM de Tris HCl pH 7.5, 50 mM NaCl, 0.5 mM de EDTA) containing GSH, GSSH, or the acetonic sulfur solution at 100 µM, for 1 h at 4 °C. Te 115 bp DNA probe (300 ng) was subsequently added to the reaction and the mixtures were further incubated at 4 °C for 30 min. BigR samples previously treated with H2S gas were also incubated 4 °C for 30 min with the DNA probe. Protein-DNA complexes were resolved on 6% polyacrylamide gels in 1X TBE bufer. Afer the run, gels were stained with ethidium bromide and visualized under UV light.

SCiENtifiC Reports | (2018)8:3508 | DOI:10.1038/s41598-018-21974-x 10 www.nature.com/scientificreports/

Mass spectrometry analyses. Recombinant BigR (50 µg) incubated with gaseous H2S or which had been treated with either the 2.0% polysulfde solution (v/v), 125 µM GSH, 5 mM hydrogen peroxide or 5 mM dimedone in phosphate-saline (PBS) bufer for 20 min at 4 °C, were resolved on SDS-gels. BigR bands were removed and digested with chymotrypsin (1 µg) for 16 h at 37 °C. Te samples were reduced or not with DTT and/or alkylated with iodoacetamide for mass spectrometry analysis as described by Villén and Gygi57. Te peptides derived from the chymotrypsin-digested samples were desalted by stage-tips, dried in a vacuum concentrator and reconstituted in 50 µL of 0.1% formic acid. BigR peptides mixtures incubated with gaseous H2S or which had been treated with either the 2.0% polysulfde solution (v/v), 125 µM GSH, 5 mM hydrogen peroxide were separated by C18 (100 µm x 100 mm) RP-nanoUPLC (nanoAcquity, Waters) coupled with a Q-Tof Premier mass spectrometer (Waters) with nanoelectrospray source at a fow rate of 0.6 µl/min. Te gradient was 2–30% acetonitrile in 0.1% formic acid over 30 min for the digested proteins. Te nanoelectrospray voltage was set to 3.5 kV, a cone voltage of 30 V and the source temperature was 80 °C. Te instrument was operated in the ‘top three’ mode, in which one MS spectrum is acquired followed by MS/MS of the top three most-intense peaks detected. Afer MS/MS fragmentation, the ion was placed on exclusion list for 60 s and for the analysis of endogenous cleavage peptides, a real-time exclusion was used. Te spectra were acquired using sofware MassLynx v.4.1 and the raw data fles were converted to a peak list format (mgf) by the sofware Mascot Distiller v.2.6.2.0, 2017 (Matrix Science Ldt.) and searched against a customized Xylella database (6,493 sequences; 2,025,062 residues) using Mascot engine v.2.3.0 (Matrix Science Ltd.), with carbamidomethylation (+57 Da) as fxed or variable modifcation, and oxidation of methionine (+16 Da), polysulfde or persulfde (+32, +64, +96 and +128 Da) in cysteines, considering the additional loss of one hydrogen as variable modifcations. One chymotrypsin missed cleavage and a tolerance of 0.1 Da for both precursor and fragment ions were considered. For disulfde bonds analysis, a peak list (mgf) generated by Mascot Distiller v.2.6.2.0, 2017 (Matrix Science Ldt.) was analyzed in MassMatrix sofware58 to automatically search Cys-Cys disulfde bonds against database containing the BigR protein sequence, according to the sofware instructions. Te parameters for disulfde bonds analysis used in MassMatrix sofware were carbamidomethylation (+57.021 Da), oxidation of methionine (+15.995 Da), polysulfde or persulfde (+32, +64, +96 and +128 Da) in cysteines, considering the additional loss of one hydrogen as variable modifcations, non-cleavable by enzymes, two chymotrypsin missed cleavages, and a tolerance of 0.1 Da for precursor and for fragment ions. BigR peptides mixtures treated with 5 mM dimedone in phosphate-saline (PBS) bufer were analyzed on a ETD enabled LTQ Orbitrap Velos mass spectrometer (Termo Fisher Scientifc) coupled with LC-MS/MS by an EASY-nLC system (Proxeon Biosystem). Peptides were separated by a 2–30% acetonitrile gradient in 0.1% formic acid using an analytical PicoFrit ID75 μm column (New Objective) at a fow rate of 300 nL/min over 30 min. Te nanoelectrospray voltage, source temperature and instrument method were set to 2.2 kV, 275 °C and data-dependent acquisition mode. Te full scan MS spectra (m/z 300-1,600) were acquired in the Orbitrap ana- lyzer afer accumulation to a target value of 1e6. Te resolution in the Orbitrap system was set to r = 60,000, and the fve most intense peptide ions with charge states ≥2 were sequentially isolated to a target value of 50,000 and fragmented in higher-energy collisional dissociation with normalized collision energy of 40% with the resolution in the Orbitrap system was set to r = 7,500 for MS/MS. Te signal threshold for triggering an MS/MS event was set to 80,000 counts and an activation time of 0.1 ms was used. Dynamic exclusion was enabled with an exclusion size list of 400, exclusion duration of 60 s, and a repeat count of 2. Te raw data fles were converted to a peak list format (mgf) by the sofware Mascot Distiller v.2.6.2.0, 2017 (Matrix Science Ldt.) using the same parameters described above with a tolerance of 10 ppm for precursor ions and 0.02 Da for fragment ions.

Reporter gene assays. GFP fuorescence, as a measure of operon activation, was determined in the C58 A. tumefaciens wild type, blh− and bigR− cells based on Barbosa and Benedetti20. Bacterial cells were grown in YEP medium for 16 h at 30 °C under agitation (200 rpm). Te cells were pelleted and washed twice in 25 mM phosphate bufer, pH 7.4, containing 0.6% glucose. Te cells were suspended in the same bufer and the optical density of the suspensions at 600 nm was adjusted to 1.0. Te cells were incubated at 30 °C for 3 h at 100 rpm in the presence or absence of the polysulfde solution at 0.2%, 2.0% or 10% v/v. Afer the incubation, bacterial cells were visualized for GFP fuorescence on a Zeiss Imager A.2 fuorescence microscope. References 1. Paul, B. D. & Snyder, S. H. H2S signaling through protein sulfydration and beyond. Nat. Rev. Mol. Cell Biol. 13, 499–407 (2012). 2. Stein, A. & Bailey, S. M. Redox biology of hydrogen sulfde: implications for physiology, pathophysiology, and pharmacology. Redox Biol. 1, 32–39 (2013). 3. Li, Z. G., Min, X. & Zhou, Z. H. Hydrogen sulfde: a signal molecule in plant cross-adaptation. Front. Plant Sci. 7, 1621 (2016). 4. Clarke, P. H. Hydrogen sulphide production by bacteria. J. Gen. Microbiol. 8, 397–407 (1953). 5. Barrett, E. L. & Clark, M. A. Tetrathionate reduction and production of hydrogen sulfde from thiosulfate. Microbiol. Rev. 51, 192–205 (1987). 6. Shimada, T., Tanaka, K. & Ishihama, A. Transcription factor DecR (YbaO) controls detoxifcation of L-cysteine in Escherichia coli. Microbiology 162, 1698–1707 (2016). 7. Lloyd, D. Hydrogen sulfde: clandestine microbial messenger? Trends Microbiol. 14, 456–462 (2006). 8. Shatalin, K., Shatalina, E., Mironov, A. & Nudler, E. H2S: a universal defense against antibiotics in bacteria. Science 334, 986–990 (2011). 9. Wu, G., Wan, F., Fu, H., Li, N. & Gao, H. A matter of timing: contrasting efects of hydrogen sulfde on oxidative stress response in Shewanella oneidensis. J. Bacteriol. 197, 3563–3572 (2015). 10. Rohwerder, T. & Sand, W. The sulfane sulfur of persulfides is the actual substrate of the sulfur-oxidizing enzymes from Acidithiobacillus and Acidiphilium spp. Microbiology 149, 1699–1710 (2003). 11. Hildebrandt, T. M. & Grieshaber, M. K. Tree enzymatic activities catalyze the oxidation of sulfde to thiosulfate in mammalian and invertebrate mitochondria. FEBS J. 275, 3352–3361 (2008). 12. Tiranti, V. et al. Loss of ETHE1, a mitochondrial dioxygenase, causes fatal sulfde toxicity in ethylmalonic encephalopathy. Nat. Med. 15, 200–205 (2009).

SCiENtifiC Reports | (2018)8:3508 | DOI:10.1038/s41598-018-21974-x 11 www.nature.com/scientificreports/

13. Guimarães, B. G. et al. Plant pathogenic bacteria utilize bioflm growth-associated repressor (BigR), a novel winged-helix redox switch, to control hydrogen sulfde detoxifcation under hypoxia. J. Biol. Chem. 286, 26148–26157 (2011). 14. Melideo, S. L., Jackson, M. R. & Jorns, M. S. Biosynthesis of a central intermediate in hydrogen sulfde metabolism by a novel human sulfurtransferase and its yeast ortholog. Biochemistry 53, 4739–4753 (2014). 15. Shen, J. et al. Staphylococcus aureus CstB is a novel multidomain persulfde dioxygenase-sulfurtransferase involved in hydrogen sulfde detoxifcation. Biochemistry 54, 4542–4554 (2015). 16. Birke, H., Hildebrandt, T. M., Wirtz, M. & Hell, R. Sulfde detoxifcation in plant mitochondria. Methods Enzymol. 555, 271–286 (2015). 17. Viscomi, C. et al. Combined treatment with oral metronidazole and N-acetylcysteine is efective in ethylmalonic encephalopathy. Nat Med. 16, 869–871 (2010). 18. Zhang, X., Vincent, A. S., Halliwell, B. & Wong, K. P. A mechanism of sulfite neurotoxicity: direct inhibition of glutamate dehydrogenase. J. Biol. Chem. 279, 43035–43045 (2004). 19. Niknahad., H. & O’Brien, P. J. Mechanism of sulfte cytotoxicity in isolated rat hepatocytes. Chem Biol Interact. 174, 147–154 (2008). 20. Barbosa, R. L. & Benedetti, C. E. BigR, a transcriptional repressor from plant-associated bacteria, regulates an operon implicated in bioflm growth. J. Bacteriol. 189, 6185–6194 (2007). 21. Pettinati, I., Brem, J., McDonough, M. A. & Schofeld, C. J. Crystal structure of human persulfde dioxygenase: structural basis of ethylmalonic encephalopathy. Hum Mol Genet. 24, 2458–2469 (2015). 22. Hänzelmann, P. et al. Crystal structure of YnjE from Escherichia coli, a sulfurtransferase with three rhodanese domains. Protein Sci. 18, 2480–2491 (2009). 23. Banerjee, R. et al. Assay methods for H2S biogenesis and catabolism enzymes. Methods Enzymol. 554, 189–200 (2015). 24. Wagner, B. A., Venkataraman, S. & Buettner, G. R. Te rate of oxygen utilization by cells. Free Radic. Biol. Med. 51, 700–712 (2011). 25. Nandi, P. K., Agrawal, M., Agrawal, S. B. & Rao, D. N. Physiological responses of Vicia faba plants to sulfur dioxide. Ecotoxicol Environ Saf. 19, 64–71 (1990). 26. Brychkova, G. et al. Sulfte oxidase protects plants against sulfur dioxide toxicity. Plant J. 50, 696–709 (2007). 27. Yarmolinsky, D., Brychkova, G., Fluhr, R. & Sagi, M. Sulfte reductase protects plants against sulfte toxicity. Plant Physiol. 161, 725–743 (2013). 28. Brychkova, G., Grishkevich, V., Fluhr, R. & Sagi, M. An essential role for tomato sulfte oxidase and enzymes of the sulfte network in maintaining leaf sulfte homeostasis. Plant Physiol. 161, 148–64 (2013). 29. Luebke, J. L., Arnold, R. J. & Giedroc, D. P. Selenite and tellurite form mixed seleno- and tellurotrisulfides with CstR from Staphylococcus aureus. Metallomics. 5, 335–342 (2013). 30. Luebke, J. L. et al. Te CsoR-like sulfurtransferase repressor (CstR) is a persulfde sensor in Staphylococcus aureus. Mol Microbiol. 94, 1343–1360 (2014). 31. Shimizu, T. et al. Sulfide-responsive transcriptional repressor SqrR functions as a master regulator of sulfide-dependent photosynthesis. Proc. Natl. Acad. Sci. USA 114, 2355–2360 (2017). 32. Debiemme-Chouvy, C., Wartelle, C. & Sauvage, F. X. First evidence of the oxidation and regeneration of polysulfdes at a GaAs electrode, under anodic conditions. A study by in situ UV-visible spectroelectrochemistry. J. Phys. Chem. B. 108, 18291–18296 (2004). 33. Greiner, R. et al. Polysulfdes link H2S to protein thiol oxidation. Antioxid. Redox Signal. 19, 1749–1765 (2013). 34. Johansson, M. & Lundberg, M. Glutathionylation of beta-actin via a cysteinyl sulfenic acid intermediary. BMC Biochem. 8, 26 (2007). 35. Rehder, D. S. & Borges, C. R. Cysteine sulfenic acid as an intermediate in disulfde bond formation and nonenzymatic protein folding. Biochemistry 49, 7748–7755 (2010). 36. Gupta, V. & Carroll, K. S. Sulfenic acid chemistry, detection and cellular lifetime. Biochim. Biophys Acta. 1840, 847–875 (2014). 37. Cuevasanta, E. et al. Reaction of hydrogen sulfde with disulfde and sulfenic acid to form the strongly nucleophilic persulfde. J. Biol. Chem. 290, 26866–80 (2015). 38. Mukherjee, D., Datta, A. B. & Chakrabarti, P. Crystal structure of HlyU, the hemolysin gene transcription activator, from Vibrio cholerae N16961 and functional implications. Biochim. Biophys Acta. 1844, 2346–2354 (2014). 39. Bhattacharyya, A. et al. Draf sequencing and comparative genomics of Xylella fastidiosa strains reveal novel biological insights. Genome Res. 12, 1556–1563 (2002). 40. Tomson, C. J. & Greenway, H. Metabolic evidence for stelar anoxia in maize roots exposed to low o(2) concentrations. Plant Physiol. 96, 1294–1301 (1991). 41. van Dongen, J. T., Schurr, U., Pfster, M. & Geigenberger, P. Phloem metabolism and function have to cope with low internal oxygen. Plant Physiol. 131, 1529–1543 (2003). 42. Armstrong, W., Webb, T., Darwent, M. & Beckett, P. M. Measuring and interpreting respiratory critical oxygen pressures in roots. Ann. Bot. 103, 281–283 (2009). 43. Krishnan, N., Fu, C., Pappin, D. J. & Tonks, N. K. H2S-induced sulfhydration of the phosphatase PTP1B and its role in the endoplasmic reticulum stress response. Sci. Signal. 4, ra86 (2011). 44. Chung, H. S., Wang, S. B., Venkatraman, V., Murray, C. I. & Van Eyk, J. E. Cysteine oxidative posttranslational modifcations: emerging regulation in the cardiovascular system. Circ. Res. 112, 382–392 (2013). 45. Sen, N. et al. Hydrogen sulfde-linked sulfydration of NF-κB mediates its antiapoptotic actions. Mol Cell. 45, 13–24 (2012). 46. Studier, F. W. Protein production by auto-induction in high-density shaking cultures. Prot. Exp. Pur. 41, 207–234 (2005). 47. Leinweber, F. J. & Monty, K. J. Sulfte determination: fuchsin method. Methods Enzymol. 143, 15–17 (1987). 48. Balan, A. et al. Te molybdate-binding protein (ModA) of the plant pathogen Xanthomonas axonopodis pv. citri. Protein Expr. Purif. 50, 215–222 (2006). 49. Solé, V. A., Papillon, E., Cotte, M., Walter, P. & Susini, J. A multiplatform code for the analysis of energy-dispersive X-ray fuorescence spectra. Spectrochim. Acta Part B 62, 63–68 (2007). 50. Chilton, M. D. et al. Agrobacterium tumefaciens DNA and PS8 bacteriophage DNA not detected in crown gall tumors. Proc. Natl. Acad. Sci. USA 71, 3672–3676 (1974). 51. Almeida, R. P. P., Pereira, E. F., Purcell, A. H. & Lopes, J. R. S. Multiplication and movement of a citrus strain of Xylella fastidiosa within sweet orange. Plant Dis. 85, 382–386 (2001). 52. Oliveira, A. C. et al. Quantifcation of Xylella fastidiosa from citrus trees by real-time polymerase chain reaction assay. Phytopathology 92, 1048–1054 (2002). 53. Mafra, V. et al. Reference genes for accurate transcript normalization in citrus genotypes under diferent experimental conditions. PLoS One 7, e31263 (2012). 54. de Oliveira, M. L. et al. Increased resistance against citrus canker mediated by a citrus mitogen-activated protein kinase. Mol. Plant Microbe Interact. 26, 1190–1199 (2013). 55. Brychkova, G., Yarmolinsky, D., Fluhr, R. & Sagi, M. Te determination of sulfte levels and its oxidation in plant leaves. Plant Sci. 190, 123–130 (2012). 56. Kleinjan, W. E., de Keizer, A. & Janssen, A. J. Kinetics of the chemical oxidation of polysulfde anions in aqueous solution. Water Res. 39, 4093–4100 (2005).

SCiENtifiC Reports | (2018)8:3508 | DOI:10.1038/s41598-018-21974-x 12 www.nature.com/scientificreports/

57. Villén, J. & Gygi, S. P. Te SCX/IMAC enrichment approach for global phosphorylation analysis by mass spectrometry. Nat. Protoc. 3, 1630–1638 (2008). 58. Xu, H. & Freitas, M. A. A mass accuracy sensitive probability based scoring algorithm for database searching of tandem mass spectrometry data. BMC Bioinformatics. 8, 133 (2007). Acknowledgements We thank Jaqueline Silva and Jackeline Zanella for helping with protein expression and purifcation, and Romenia Domingues for helping with the mass spectrometry data acquisition. We also thank Fabio Zambello for helping with the hydrogen sulfde experiments performed at the Brazilian Synchrotron Light Laboratory (LNLS). Tis work was supported by Fundação de Amparo à Pesquisa do Estado de São Paulo - FAPESP (grant 2011/20468- 1). CEB and NPVL received a fellowship from Conselho Nacional de Desenvolvimento Científco e Tecnológico (CNPq) and FAPESP, respectively. Author Contributions C.E.B. conceived and coordinated the study and wrote the paper. N.P.V.L. designed, performed and analyzed the experiments shown in all Figures. N.P.V.L., B.A.P. and A.F.P.L. designed, performed and analyzed the mass spectrometry experiments. A.C.M. and A.E.V. provided technical assistance and contributed to the experiments shown in Figure 4. C.A.P. designed, performed and analyzed the X-ray fuorescence experiments. R.C. and A.A.S. performed plant inoculations and bacterial quantifcation in citrus. All authors reviewed the results and approved the fnal version of the manuscript. Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-018-21974-x. Competing Interests: Te authors declare no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. Te images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

SCiENtifiC Reports | (2018)8:3508 | DOI:10.1038/s41598-018-21974-x 13